A Comprehensive Benchmark Dataset for Agricultural Field Boundaries
Fields of The World (FTW) is a comprehensive benchmark dataset designed to enhance the development of machine learning models for instance segmentation of agricultural field boundaries. It aggregates and harmonizes a number of open datasets into 1.6 million parcel boundaries and over 70,000 samples covering diverse agricultural landscapes across 4 continents and 27 countries.
The extensive geographic coverage and large amount of data allows for the development of models that can generalize well to different agricultural practices and field types.
The dataset is well-structured with great documentation and tools so that any AI/ML researcher can immediately experiment and build models, even without geospatial expertise.
FTW is built with a number of open building blocks so that it is easy for a variety of individuals and organizations to contribute their data, software, compute, and expertise.
There is a growing need for accurate and scalable field boundary data for global agricultural monitoring and assessments. Traditional data collection has not met that need, but there is large potential to use AI & Earth Observation Imagery to meet the demand. Fields of the World aims to be an "ImageNet for Fields" by creating a foundational dataset that spurs on a greater ecosystem of collaboration.
We welcome contributions from the community! Our goal is to create an open, collaborative ecosystem where we all work together to advance a commons of data, models, benchmarks and software for agricultural insights.